FLOOD00113

• County: Statewide
• Title: Changes in the Hydrological Response of a Small Mediterranean Basin a Year After a Wildfire
• Date: 6/8/1992
• Prepared For: Journal of Hydrology
• Prepared By: Elsevier Science Publishers
• Floodplain - Doc Type: Educational/Technical/Reference Information

298

J. LAVABRE ET AL.

HYDROLOGICAL RESPONSE OF A MEDITERRANEAN BASIN

299

IIIF P' (0.
E" -( (S/A) {2 - (51 A)))

(1- (51...,) 2) P'

t_~
..................
A I ................................... I
; :jjj:jt:ji::j~j~~:j:j~jj~;~:~~:;

l
~

(5/,,)2 P'

~~ t
16
I
............ ....
I ..R.............................. I
, ::..::::::::::::::::::::::::::::: ,

~'
a.-
R-6

where B is the maximal storage.
This scheme, repeated each time step, is used with the monthly version.
However, to fit better the hydrographs observed at daily time step it is
necessary to add both a third module and a third parameter.
At daily and lower time steps it is necessary to introduce a time delay to the
peak since rainfall does not usually have an immediate response. To incor-
porate this time delay, a delay function, rather similar to the isochrone notion,
is added to the model. It has an imposed parabolic structure, and the only
parameter to supply is the total time delay, C. since its integral must be unity
to preserve the volume conservation (see Editjano and Michel, 1989). The
parameters A and B (plus C in the case of the daily simulation) are optimised
using a gradient method by minimisation of the sum of the squared residuals
between observed and simulated runoff.
Finally some recent improvements have been incorpo~ated in the daily
version (Editjano et aI., 1991). As for the routing reservoir emptying, a general
power function instead of the previous quadratic function is used. The new
power used was five instead of two, allowing the recession curve to fit better
than previously. Additionally, after the separation of the term that will
produce runoff, only 90% of it (0.9 (SIA)' P') is now delayed and routed, and
the other 10% is directly used to give rise to the outflow. Thus Q at each time
step will be the addition of two terms: the 10% of the available water at this
step, (0.1 (SIA)' P'), and the outflow produced by the routing reservoir which
is in fact the result of the previous rainfall. Editjano et a!. (1991) found that
this modification allows the model to fit quick runoff responses beller.

IIF P' > 0 t

Fig. AI. Diagram of the OR3 rainfall-runoff model.

soil moisture reservoir is emptying as (Editjano and Michel, 1989)

dS = - [~( 2 - ~)] dE'

where E' is E - P. Therefore, there is no water available to produce runoff
and the outflow is reduced to the routing reservoir emptying;
(2) if P' > 0, there are no evapotranspiration losses from the soil moisture
reservoir, and the water available P' will be divided into two terms: [I - (SI
A)2] P' will be retained by the soil moisture reservoir increasing its storage
level, S, and the complementary portion, (SIA)' P' will refill the routing
reservoir.
As ror the routing reservoir, we will consider that this emptying in absence
of refill is a quadratic relation of the storage level R (Editjano and Michel,
1989)

dR _ kR'
dI-

So the outflow durin& a time step is given by

R'
Q=R+B